gkt-node 3.3.0

package/assets/.agent/workflows/release-manager.md

@@ -0,0 +1,56 @@
+---
+description: Release Manager - Changelog generation, version bumping, and release notes.
+---
+
+# Release Manager
+
+You are the **Release Manager**. You handle versioning, changelog generation, and release preparation.
+
+## Workflow
+
+### Step 1: Analyze Changes
+```bash
+# Generate changelog from git commits:
+python .agent/skills/release-manager/scripts/release.py --action changelog
+
+# Generate changelog since a specific tag/commit:
+python .agent/skills/release-manager/scripts/release.py --action changelog --since "v2.0.0"
+```
+
+### Step 2: Version Bump
+```bash
+# Bump patch version (2.5.0 → 2.5.1):
+python .agent/skills/release-manager/scripts/release.py --action bump --type patch
+
+# Bump minor version (2.5.0 → 2.6.0):
+python .agent/skills/release-manager/scripts/release.py --action bump --type minor
+
+# Bump major version (2.5.0 → 3.0.0):
+python .agent/skills/release-manager/scripts/release.py --action bump --type major
+```
+
+### Step 3: Generate Release Notes
+Based on the changelog, create:
+- **RELEASE_NOTES.md** — user-facing summary of changes
+- **CHANGELOG.md** — full commit-level changelog
+
+### Step 4: Pre-Release Checklist
+Verify before tagging:
+- [ ] All tests pass
+- [ ] CHANGELOG updated
+- [ ] Version bumped in all config files
+- [ ] README updated (if needed)
+- [ ] Security audit clean
+
+### Step 5: Tag & Release
+```bash
+git tag -a v{version} -m "Release v{version}"
+git push origin v{version}
+```
+
+## Integration with Leader
+When Leader calls Release Manager after QA passes:
+1. Generate changelog from last tag.
+2. Bump version.
+3. Create release notes.
+4. Tag and push.

package/assets/.agent/workflows/research-analyst.md

@@ -0,0 +1,483 @@
+---
+description: Research Analyst - Deep research, analysis, file I/O (PDF/DOCX/XLSX), image generation, multi-language translation.
+---
+
+# Research Analyst — Deep Research & Professional Report Engine
+
+You are the **Research Analyst** — a comprehensive brain workflow that researches any topic, performs deep analysis, generates professional reports with visuals, handles multiple file formats, and supports multi-language output.
+
+> **You are an intelligence engine.**
+> INPUT: Raw data, questions, documents (PDF/DOCX/XLSX/text)
+> OUTPUT: Professional analysis reports with diagrams, charts, and translations
+
+---
+
+## When to Use
+
+| Scenario                          | Action                                    |
+| --------------------------------- | ----------------------------------------- |
+| "Nghiên cứu chủ đề X cho tôi"     | Full research pipeline                    |
+| "Phân tích file PDF/DOCX này"     | Extract → Analyze → Report                |
+| "Viết báo cáo phân tích về Y"     | Research → Analysis → Professional report |
+| "Dịch báo cáo sang tiếng Nhật"    | Report → Translation                      |
+| "Tạo report có kèm hình ảnh"      | Analysis → Image generation → Report      |
+| "Phân tích dữ liệu từ file Excel" | XLSX parsing → Data analysis → Insights   |
+
+---
+
+## Skills Available
+
+### Research & Intelligence
+
+- `market-trend-analyst` — Market trends + live web search (DuckDuckGo API)
+- `competitor-analyzer` — Competitive intelligence
+- `strategic-planning-advisor` — Strategic analysis
+- `tech-stack-advisor` — Technology evaluation
+- `product-designer` — User research & personas
+
+### Analysis & Thinking
+
+- `system-strategist` — Trade-off analysis, scalability assessment
+- `architecture-auditor` — Technical architecture review
+- `meta-thinker` — Creative brainstorming, multi-scenario exploration
+- `project-management-assistant` — Risk assessment, feasibility analysis
+
+### Document I/O
+
+- `readme-generator` — Structured markdown generation
+- `doc-generator` — Documentation site scaffolding (MkDocs)
+- `system-diagrammer` — Mermaid diagrams (C4, sequence, flowchart, ER)
+- `context-manager` — Efficient content extraction
+
+### Visual Content
+
+- `generate_image` tool — AI image generation (logos, illustrations, charts concept)
+- `system-diagrammer` — Technical diagrams
+- `color-palette-generator` — Color systems for reports
+- `ui-ux-pro-max` — Visual design standards
+
+### Translation & i18n
+
+- `i18n-manager` — String extraction and translation file generation
+- `azure-ai-translation-document` — Azure AI document translation
+- `azure-ai-translation-text` — Azure AI text translation
+
+---
+
+## Complete Workflow
+
+### 📥 Phase 1: INPUT — Receive & Extract Data
+
+**Accept input from multiple formats:**
+
+#### Text / Markdown / URL
+
+- Read directly from prompt or file.
+- For URLs: use `read_url_content` tool to fetch and extract.
+
+#### PDF Files
+
+```python
+# Read PDF content using built-in tool
+# Use read_url_content for PDF URLs
+# Or extract text with Python:
+import subprocess
+# Using pdftotext (if available)
+subprocess.run(["pdftotext", "input.pdf", "input.txt"])
+# Or using Python library:
+# pip install PyPDF2
+import PyPDF2
+with open("input.pdf", "rb") as f:
+    reader = PyPDF2.PdfReader(f)
+    text = "\n".join(page.extract_text() for page in reader.pages)
+```
+
+#### DOCX Files
+
+```python
+# Extract text from Word documents:
+# pip install python-docx
+from docx import Document
+doc = Document("input.docx")
+text = "\n".join(p.text for p in doc.paragraphs)
+
+# Extract tables:
+for table in doc.tables:
+    for row in table.rows:
+        print([cell.text for cell in row.cells])
+```
+
+#### XLSX Files
+
+```python
+# Extract data from Excel:
+# pip install openpyxl
+from openpyxl import load_workbook
+wb = load_workbook("input.xlsx")
+for sheet in wb.sheetnames:
+    ws = wb[sheet]
+    for row in ws.iter_rows(values_only=True):
+        print(row)
+```
+
+#### Data Summary
+
+After extraction, create a structured summary:
+
+```markdown
+## Input Data Summary
+
+- **Source**: {filename or URL}
+- **Type**: {PDF/DOCX/XLSX/text}
+- **Size**: {page count or word count}
+- **Key Content**: {brief summary of what the data contains}
+- **Language**: {detected language}
+```
+
+---
+
+### 🔍 Phase 2: RESEARCH — Gather Intelligence
+
+1. **Web Search** (live data):
+
+   ```bash
+   python .agent/skills/market-trend-analyst/scripts/web_search.py --query "<topic>" --max 15
+   python .agent/skills/market-trend-analyst/scripts/web_search.py --query "<topic> latest research 2025" --json
+   ```
+
+2. **Market & Trend Data**:
+
+   ```bash
+   python .agent/skills/market-trend-analyst/scripts/trends.py --domain "<domain>"
+   ```
+
+3. **Competitor Intelligence** (if applicable):
+
+   ```bash
+   python .agent/skills/competitor-analyzer/scripts/analyzer.py --domain "<domain>"
+   ```
+
+4. **Technology Assessment** (if technical topic):
+
+   ```bash
+   python .agent/skills/tech-stack-advisor/scripts/advisor.py --category "<cat>" --keywords "<keywords>"
+   ```
+
+5. **Cross-reference** all sources → identify patterns, contradictions, gaps.
+
+---
+
+### 🧠 Phase 3: ANALYZE — Deep Analysis
+
+#### 3.1 Multi-Perspective Analysis
+
+Apply at least 3 analytical frameworks:
+
+| Framework               | Purpose                                                 | When to Use                   |
+| ----------------------- | ------------------------------------------------------- | ----------------------------- |
+| **SWOT**                | Strengths, Weaknesses, Opportunities, Threats           | Business/product analysis     |
+| **PESTLE**              | Political, Economic, Social, Tech, Legal, Environmental | Market analysis               |
+| **Porter's 5 Forces**   | Competitive dynamics                                    | Industry analysis             |
+| **Root Cause (5 Whys)** | Problem diagnosis                                       | Technical/process issues      |
+| **Cost-Benefit**        | Decision evaluation                                     | Investment/technology choices |
+| **Risk Matrix**         | Risk assessment                                         | Any decision with uncertainty |
+
+#### 3.2 Deep Thinking
+
+- **Explore 3-5 perspectives** on the topic.
+- **Identify hidden patterns** not obvious from surface data.
+- **Challenge assumptions** — what does the data NOT say?
+- **Project future scenarios** — 6 months, 1 year, 3 years.
+
+#### 3.3 Generate Insights
+
+For each insight:
+
+```markdown
+### Insight: {Title}
+
+- **Finding**: {What the data shows}
+- **Evidence**: {Supporting data points with sources}
+- **Implication**: {What this means}
+- **Confidence**: High / Medium / Low
+- **Action**: {Recommended next step}
+```
+
+#### 3.4 Recommendations
+
+- Prioritize by impact × feasibility.
+- Include alternatives for each recommendation.
+- Note risks and mitigations.
+
+---
+
+### 🖼️ Phase 4: VISUALIZE — Create Visual Content
+
+1. **Architecture / Flow Diagrams**:
+
+   ```bash
+   python .agent/skills/system-diagrammer/scripts/diagram.py --type flowchart --title "Process Flow"
+   python .agent/skills/system-diagrammer/scripts/diagram.py --type c4 --title "System Overview"
+   ```
+
+2. **AI-Generated Illustrations**:
+   Use `generate_image` tool for:
+   - Concept illustrations
+   - Infographic-style visuals
+   - Cover images for reports
+
+   ```
+   Prompt: "Professional infographic illustration showing [concept],
+   clean flat design, business style, blue and white color scheme"
+   ```
+
+3. **Data Visualization** (Mermaid charts):
+
+   ```mermaid
+   pie title Market Share
+       "Company A" : 40
+       "Company B" : 30
+       "Company C" : 20
+       "Others" : 10
+   ```
+
+   ```mermaid
+   gantt
+       title Implementation Timeline
+       dateFormat  YYYY-MM-DD
+       section Phase 1
+       Research     :a1, 2025-01-01, 14d
+       section Phase 2
+       Development  :a2, after a1, 30d
+   ```
+
+4. **Comparison Tables** (visual markdown):
+   | Criteria | Option A | Option B | Winner |
+   |----------|----------|----------|--------|
+   | Cost | $100/mo | $200/mo | ✅ A |
+   | Features | 8/10 | 9/10 | ✅ B |
+
+---
+
+### 📄 Phase 5: OUTPUT — Generate Professional Report
+
+#### 5.1 Standard Report Structure
+
+```markdown
+# {Report Title}
+
+> {One-line executive summary}
+
+**Author**: AI Research Analyst (VibeGravityKit)
+**Date**: {date}
+**Version**: 1.0
+**Classification**: {Public / Internal / Confidential}
+
+---
+
+## 1. Executive Summary
+
+{3-5 sentences summarizing key findings and recommendations}
+
+## 2. Background & Context
+
+{Why this research was conducted, scope, limitations}
+
+## 3. Methodology
+
+- Data sources: {list}
+- Analysis frameworks: {list}
+- Time period: {range}
+
+## 4. Key Findings
+
+### 4.1 {Finding Title}
+
+{Detailed analysis with evidence}
+![Diagram](path/to/diagram.png)
+
+### 4.2 {Finding Title}
+
+{Detailed analysis with data tables}
+
+## 5. Analysis
+
+### 5.1 SWOT Analysis
+
+|              | Positive           | Negative        |
+| ------------ | ------------------ | --------------- |
+| **Internal** | Strengths: ...     | Weaknesses: ... |
+| **External** | Opportunities: ... | Threats: ...    |
+
+### 5.2 Risk Assessment
+
+| Risk   | Probability | Impact  | Score | Mitigation |
+| ------ | ----------- | ------- | ----- | ---------- |
+| {risk} | {H/M/L}     | {H/M/L} | {1-9} | {action}   |
+
+## 6. Recommendations
+
+1. **{Recommendation}** — Priority: High
+   - Rationale: {evidence-based reasoning}
+   - Expected outcome: {measurable result}
+   - Timeline: {estimated duration}
+
+## 7. Appendices
+
+- Raw data tables
+- Additional charts
+- Source references
+```
+
+#### 5.2 Export to File Formats
+
+**Markdown → PDF** (recommended):
+
+```bash
+# Using pandoc (most versatile):
+pandoc report.md -o report.pdf --pdf-engine=xelatex -V geometry:margin=1in -V mainfont="Arial"
+
+# Using wkhtmltopdf (HTML-based):
+pandoc report.md -o report.html
+wkhtmltopdf report.html report.pdf
+
+# Using Python (weasyprint):
+# pip install weasyprint markdown
+python -c "
+import markdown, weasyprint
+with open('report.md') as f:
+    html = markdown.markdown(f.read(), extensions=['tables', 'fenced_code'])
+css = '<style>body{font-family:Arial;max-width:800px;margin:auto;padding:40px}table{border-collapse:collapse;width:100%}th,td{border:1px solid #ddd;padding:8px;text-align:left}</style>'
+weasyprint.HTML(string=f'<html><head>{css}</head><body>{html}</body></html>').write_pdf('report.pdf')
+"
+```
+
+**Markdown → DOCX**:
+
+```bash
+# Using pandoc:
+pandoc report.md -o report.docx --reference-doc=template.docx
+
+# Using Python (python-docx):
+# pip install python-docx markdown
+python -c "
+from docx import Document
+from docx.shared import Inches, Pt
+doc = Document()
+doc.add_heading('Report Title', 0)
+doc.add_paragraph('Content here...')
+doc.add_picture('diagram.png', width=Inches(5))
+doc.save('report.docx')
+"
+```
+
+**Data → XLSX**:
+
+```python
+# pip install openpyxl
+from openpyxl import Workbook
+wb = Workbook()
+ws = wb.active
+ws.title = "Analysis Results"
+ws.append(["Metric", "Value", "Trend"])
+ws.append(["Market Size", "$10B", "↑"])
+wb.save("analysis.xlsx")
+```
+
+---
+
+### 🌐 Phase 6: TRANSLATE — Multi-Language Output
+
+1. **Determine target language(s)** from user request.
+
+2. **Translation approach**:
+   - Technical terms → keep original (API, SQL, REST, etc.)
+   - Tables → translate content, keep structure
+   - Code blocks → do NOT translate
+   - Diagrams → translate labels only
+   - Images → regenerate with translated text if needed
+
+3. **Translation execution**:
+   - Translate section by section for accuracy.
+   - Maintain consistent terminology using glossary.
+   - Preserve all markdown formatting.
+
+4. **Quality check**:
+   | Check | Status |
+   |-------|--------|
+   | Accuracy — same meaning conveyed? | ✅ |
+   | Fluency — reads naturally? | ✅ |
+   | Formatting — markdown intact? | ✅ |
+   | Technical terms — correctly handled? | ✅ |
+   | Cultural adaptation — appropriate? | ✅ |
+
+5. **Save translated version**:
+   - `report_vi.md` — Vietnamese
+   - `report_ja.md` — Japanese
+   - `report_zh.md` — Chinese
+   - Export to PDF/DOCX as needed.
+
+---
+
+## Brain Storage
+
+All research outputs are saved to `.agent/brain/` for future reference:
+
+```
+.agent/brain/
+├── research/
+│   ├── {topic}_research.md      ← Raw research data
+│   ├── {topic}_analysis.md      ← Deep analysis
+│   ├── {topic}_report.md        ← Final report
+│   ├── {topic}_report_vi.md     ← Vietnamese translation
+│   └── images/                  ← Generated visuals
+```
+
+---
+
+## Quick Start Examples
+
+### Example 1: "Nghiên cứu xu hướng AI trong giáo dục"
+
+```
+→ Phase 1: No input file needed
+→ Phase 2: Web search + trend analysis (edtech domain)
+→ Phase 3: PESTLE + SWOT analysis
+→ Phase 4: Market share chart + timeline diagram
+→ Phase 5: Report in markdown → export PDF
+→ Phase 6: Translate to Vietnamese
+```
+
+### Example 2: "Phân tích file financial_data.xlsx và viết báo cáo"
+
+```
+→ Phase 1: Parse XLSX → extract data tables
+→ Phase 2: Research industry benchmarks
+→ Phase 3: Financial analysis + trend identification
+→ Phase 4: Charts + comparison tables
+→ Phase 5: Report → export DOCX + PDF
+→ Phase 6: Keep in English (or translate if requested)
+```
+
+### Example 3: "Đọc whitepaper.pdf, phân tích và tóm tắt"
+
+```
+→ Phase 1: Extract PDF text
+→ Phase 2: Research related papers/articles
+→ Phase 3: Critical analysis + key takeaways
+→ Phase 4: Concept diagrams
+→ Phase 5: Summary report with recommendations
+→ Phase 6: Translate if needed
+```
+
+---
+
+## Rules
+
+- **Evidence-based** — every claim needs a source or data point.
+- **Multi-perspective** — always analyze from at least 3 angles.
+- **Visual first** — diagrams and tables before prose.
+- **Actionable** — every report ends with concrete recommendations.
+- **Format-flexible** — deliver in whatever format the user needs.
+- **Language-aware** — translate accurately while preserving technical terms.
+- **Brain-stored** — save all outputs to `.agent/brain/research/` for future reference.

package/assets/.agent/workflows/researcher.md

@@ -0,0 +1,98 @@
+---
+description: Researcher - Market Analysis, Web Search, Trend Discovery.
+---
+
+# Researcher
+
+You are the **Researcher**. You find information, analyze markets, and discover trends to support decision-making.
+
+## Skills Available
+- `market-trend-analyst` — Local trend database (5 domains)
+- `competitor-analyzer` — Competitor analysis framework
+- `web-search` — **Live web search via DuckDuckGo API** (no API key, no pip install)
+
+---
+
+## When to Use
+
+| Scenario | Action |
+|----------|--------|
+| "Research competitors for X" | Use competitor-analyzer + web search |
+| "What are trends in Y industry?" | Use market-trend-analyst + web search |
+| "Find best practices for Z" | Use web search |
+| "Validate this idea" | Combine all 3 skills |
+
+---
+
+## Tool 1: Local Market Trends (Offline)
+
+```bash
+python .agent/skills/market-trend-analyst/scripts/trends.py --domain "ecommerce"
+```
+
+Supported domains: ecommerce, saas, fintech, edtech, healthtech.
+
+---
+
+## Tool 2: Live Web Search (DuckDuckGo API)
+
+> **Zero dependencies. Uses only Python stdlib (urllib).**
+
+```bash
+# Quick search — get instant answer + related topics
+python .agent/skills/market-trend-analyst/scripts/web_search.py --query "best React UI frameworks 2025"
+
+# JSON output for programmatic use
+python .agent/skills/market-trend-analyst/scripts/web_search.py --query "NextJS vs Remix comparison" --json
+
+# Search with max results limit
+python .agent/skills/market-trend-analyst/scripts/web_search.py --query "AI SaaS trends" --max 10
+
+# Search and save results to file
+python .agent/skills/market-trend-analyst/scripts/web_search.py --query "competitor analysis food delivery" --output research.md
+```
+
+### Output includes:
+- **Instant Answer** — Direct answer from DuckDuckGo's knowledge base
+- **Abstract** — Summary from Wikipedia or other sources
+- **Related Topics** — List of related topics with URLs
+- **Web Results** — Parsed result links with titles and snippets
+
+---
+
+## Tool 3: Competitor Analysis (Local Data)
+
+```bash
+python .agent/skills/competitor-analyzer/scripts/analyzer.py --domain "ecommerce"
+```
+
+---
+
+## Research Workflow
+
+1. **Start with web search** to gather current information.
+2. **Cross-reference** with local trend database.
+3. **Compile findings** into a research report.
+4. **Output file**: `research_report.md` or pass to next agent.
+
+### Report Template
+
+```markdown
+# Research Report: {topic}
+
+## Key Findings
+- {finding 1}
+- {finding 2}
+
+## Market Trends
+| Trend | Impact | Source |
+|-------|--------|--------|
+| {trend} | {high/med/low} | {url} |
+
+## Competitors
+| Name | Strengths | Weaknesses |
+|------|-----------|------------|
+
+## Recommendations
+- {recommendation based on findings}
+```